A solver at the planner's service

The planner's dilemma

If you are a planner, you already know how to build a schedule. You have your spreadsheets, your rules of thumb, your years of experience. Every month, you get it done. The shifts are covered, the operation runs, everything is handled.

But you also know what it costs you. Days of manual work. Constant back-and-forth checking for conflicts. A nagging feeling that the schedule could be better, more balanced, more fair, but there is simply no time to explore every possible combination. You fix one gap, another appears. You balance one team, another loses a weekend.

The idea of adopting a new system is not appealing. You do not want to change your tools, migrate your data, or learn a new platform. Your internal procedures work. What you need is not a new system. What you need is a better result, delivered in the same format you already use, without touching anything else.

A solver that works alongside you

A constraint programming solver is not a replacement for the planner. It is a computational tool that runs in parallel with the planner's existing workflow. It takes the same data the planner already works with, applies every scheduling rule mathematically, and produces the best possible assignment. The planner gets the result, reviews it, and decides what to do with it.

Nothing changes in the planner's daily routine. The same staff data, the same shift definitions, the same operational rules. The solver simply does the computational work that no human can do manually: evaluate hundreds of thousands of combinations simultaneously and find the assignment that satisfies every constraint while optimising every objective.

Think of it as a calculator for scheduling. A planner does not stop understanding numbers because they use a calculator. They use it because it is faster, more accurate, and frees them to focus on decisions that require human judgement. The solver works the same way. It handles the combinatorial complexity. The planner handles the operational context.

Same data, same rules

The solver works with the data every planner already manages: staff lists with qualifications, shift definitions with start and end times, daily coverage requirements that follow the operational demand, and individual constraints for each employee.

This data does not need to be restructured, reformatted, or migrated. It is loaded once into a scheduling dashboard, and from that point the planner maintains it as part of their normal routine. When an agent goes on vacation, the planner records it. When staffing needs change for the winter schedule, the planner updates them. The workflow does not change.

The solver reads these parameters, builds a mathematical model, and computes the optimal assignment. The planner's HR system, payroll software, and operational tools remain completely untouched. The solver is a standalone computation layer that produces a result. It does not connect to anything else.

The result: an optimised Excel file

Every solver run produces an Excel file. A structured, readable monthly schedule with agents on rows, days on columns, and shift codes in each cell. The same format planners have always used.

This is not an approximation or a suggestion. It is the mathematically optimal assignment given the data and rules provided. The CP-SAT solver explores every valid combination and proves that no better solution exists. Every hard constraint is enforced. Every soft objective is balanced to the best achievable trade-off.

The planner opens the file, reviews it, and uses it exactly as they would use any schedule they built manually. They can print it, share it with operations, hand it to team leaders, or import it into their existing planning tools. It is a standard Excel file, not a proprietary format tied to a platform.

The planner remains the decision-maker. If they disagree with a specific assignment, they adjust it. If they want to prioritise coverage over fairness, they change the solver's weights and regenerate. The solver is a tool at their service, not a system that dictates the schedule.

Built-in compliance

When building a schedule by hand, compliance is checked after the fact. The planner scans rows and columns, looking for rest period violations, qualification mismatches, or unfair distributions. With dozens of agents and dozens of rules, errors slip through. Not because the planner lacks skill, but because the volume of combinations exceeds what any human can verify exhaustively.

A constraint solver works the other way around. Compliance is not checked after the schedule is built. It is built into the schedule from the start. Every rule is encoded as a mathematical constraint. The solver cannot produce a result that violates a hard constraint. It is structurally impossible.

Soft objectives are optimised simultaneously. The solver minimises a weighted penalty score that balances every dimension of schedule quality. The result is not just rule-compliant, it is the best schedule that can exist given the planner's data and priorities.

An optimisation score gives the planner a single number to evaluate quality. The breakdown shows exactly where trade-offs were made, giving clear guidance on what to adjust if needed.

Nothing to lose

Adopting a parallel solver carries no operational risk because it changes nothing in the existing environment.

• No system to replace: the solver runs on a separate machine, on-premise, within your network. It does not interact with your existing tools
• No procedures to change: the planner keeps their routine. The only difference is that they receive a mathematically optimised schedule instead of building one from scratch
• No learning curve: the output is an Excel file. Every planner already knows how to read one
• No dependency: if the solver is unavailable for any reason, the planner builds the schedule manually as they always have. There is no lock-in
• Side-by-side comparison: during the first months, planners can generate the optimised schedule alongside their manual version and compare. This builds confidence without any commitment

The solver runs in complete network isolation. No data leaves the planner's infrastructure. No cloud, no external API, no internet connection required.

How it starts

The setup is straightforward. The planner's existing data is loaded into the system, the scheduling rules are configured to match operational requirements, and the first optimised schedule is generated within days. The planner compares it with their manual version, provides feedback, and the model is refined until the result matches their expectations.

From that point on, generating a monthly schedule takes minutes instead of days. The solver handles the combinatorial complexity. The planner focuses on what matters: operational decisions, team management, and handling the exceptions that only human experience can resolve. Their internal procedures remain exactly the same. They simply have a better tool to do the heavy lifting.